Experimental feature postfix match

[RossSmyth]

This has a basic experimental implementation for the RFC postfix match (rust-lang/rfcs#3295, #121618). Liaison is @scottmcm with the lang team's experimental feature gate process.

This feature has had an RFC for a while, and there has been discussion on it for a while. It would probably be valuable to see it out in the field rather than continue discussing it. This feature also allows to see how popular postfix expressions like this are for the postfix macros RFC, as those will take more time to implement.

It is entirely implemented in the parser, so it should be relatively easy to remove if needed.

This PR is split in to 5 commits to ease review.

1. The implementation of the feature & gating.
2. Add a MatchKind field, fix uses, fix pretty.
3. Basic rustfmt impl, as rustfmt crashes upon seeing this syntax without a fix.
4. Add new MatchSource to HIR for Clippy & other HIR consumers

[rustbot]

r? @petrochenkov

rustbot has assigned @petrochenkov.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[rustbot]

Some changes occurred in src/tools/rustfmt

cc @rust-lang/rustfmt

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

[RossSmyth]

UI tests need to be culled

[ytmimi]

The rustfmt team prefers that support for experimental features be added in a PR to the rust-lang/rustfmt repo. I've proposed an alternative set of changes that should prevent rustfmt from removing the experimental postfix expressions, but not apply formatting to it.

[ytmimi]

Given that the postfix match is an experimental feature I'd prefer not to add formatting support in this PR. For now I think it makes sene to add just enough support so that rustfmt doesn't completely remove the postfix match.

I think the following changes would achieve that.

match kind {
    MatchKind::Prefix => rewrite_match(context, cond, arms, shape, expr.span, &expr.attrs),
    MatchKind::Postfix => {
        // experimental feature postfix match (link to RFC)
        Some(context.snippet(expr.span).to_owned())
    }
}

cc: @calebcartwright

[RossSmyth]

That is what I did originally and iirc it caused rustfmt to crash.

[ytmimi]

The only way I can think for context.snippet(expr.span).to_owned() to crash is if the expr.span isn't valid, but that seems unlikely. You could also try returning None from the MatchKind::Postfix match arm.

[calebcartwright]

I actually don't have any concerns with the formatting implementation proposed here as this is rather trivial syntactically, a reasonable initial formatting, and in accordance with the style guide policy on nightly-only/experimental syntax.

the more intriguing part for me was the mention of it "crashing", the details of which I'd be interested in seeing (though that could be shared with us separately in Zulip)

[RossSmyth]

Some sort of panic occurred. I'll try to replicate it when I work on this and post it on Zulip. Probably won't be until Friday or Saturday though.

[RossSmyth]

Ok, here's a branch with the panic: https://github.com/RossSmyth/rust/tree/rustfmt_test_panic

It be may as designed? Looking at it again after just running x.py test rustfmt

Ran 589 system tests.
assertion `left == right` failed: 1 system tests failed
  left: 1
 right: 0
Failed to join a test thread: Any { .. }


failures:
    test::system_tests

test result: FAILED. 171 passed; 2 failed; 0 ignored; 0 measured; 0 filtered out; finished in 7.15s

There are two panics that occur.

1. rust/src/tools/rustfmt/src/test/mod.rs

   Lines 367 to 369 in 279c9ba

   	// Display results.
   	println!("Ran {} idempotent tests.", count);
   	assert_eq!(fails, 0, "{} idempotent tests failed", fails);

Fails with left == right failed: 1 idempotent tests failed

3. rust/src/tools/rustfmt/src/test/mod.rs

   Lines 66 to 77 in 279c9ba

   	fn run_test_with<F>(test_setting: &TestSetting, f: F)
   	where
   	F: FnOnce(),
   	F: Send + 'static,
   	{
   	thread::Builder::new()
   	.stack_size(test_setting.stack_size)
   	.spawn(f)
   	.expect("Failed to create a test thread")
   	.join()
   	.expect("Failed to join a test thread")
   	}

Fails with panicked at src\\tools\\rustfmt\\src\\test\\mod.rs:76:10:\nFailed to join a test thread: Any { .. }

Is it designed to just panic if a test fails? It also dumps a pretty large JSON object out.

[rustbot]

Some changes occurred in match checking

cc @Nadrieril

[RossSmyth]

Two new commits:

1. Add a new MatchSource to HIR & fix up uses
2. Add basic support to clippy for postfix match

I'm not sure if the second commit is needed or desired. There is more worked needed, but as far as I can tell it doesn't break clippy. Since there are many other MatchSources, clippy guards against the other kinds well. So in the end it results in Clippy ignoring postfix match expression and not emitting any lints. A quick rg shows twenty uses of MatchSource::Normal in Clippy. Probably each needs to be looked at & possibly adjusted.

[RossSmyth]

Hmm I'm not sure why this fails in CI. It works on my computer. It looks like ui_tests isn't normalizing the test paths for some reason. Which is strange as every other Clippy UI test uses $DIR. I may just move this commit to a different branch a separate PR to Clippy. Clippy doesn't do anything bad without it.

[petrochenkov]

Hmm I'm not sure why this fails in CI. It works on my computer. It looks like ui_tests isn't normalizing the test paths for some reason. Which is strange as every other Clippy UI test uses $DIR. I may just move this commit to a different branch a separate PR to Clippy. Clippy doesn't do anything bad without it.

There are too many tests in tests/ui now, you could fix it by moving the new tests into some subdirectory, like match or parser.

[RossSmyth]

Hmm I'm not sure why this fails in CI. It works on my computer. It looks like ui_tests isn't normalizing the test paths for some reason. Which is strange as every other Clippy UI test uses $DIR. I may just move this commit to a different branch a separate PR to Clippy. Clippy doesn't do anything bad without it.

There are too many tests in tests/ui now, you could fix it by moving the new tests into some subdirectory, like match or parser.

I am aware of that. I am talking about the Clippy UI test failure.

[rustbot]

Failed to set assignee to ready: cannot assign: response: {"message":"Not Found","documentation_url":"https://docs.github.com/rest/issues/assignees#add-assignees-to-an-issue"}

Note: Only org members with at least the repository "read" role, users with write permissions, or people who have commented on the PR may be assigned.

[scottmcm]

Hello fellow lang folks! I'm not sure exactly what the process is on our side to approve this, but since it's not a one-way door let's try a poll:

@rfcbot poll T-lang Are you ok with an experimental implementation of postfix match in nightly?

I'm not planning on waiting for full checkboxes, since we can always revert the nightly implementation if necessary, but I want to give an opportunity for folks to raise objections just in case.

[rfcbot]

Team member @scottmcm has asked teams: T-lang, for consensus on:

Are you ok with an experimental implementation of postfix match in nightly?

• @joshtriplett
• @nikomatsakis
• @pnkfelix
• @scottmcm
• @tmandry

[scottmcm]

@bors r=petrochenkov

[bors]

📌 Commit 567c98b has been approved by petrochenkov

It is now in the queue for this repository.